New Feature: Measuring Pace Based on Time to Shot

Nearly five years ago, I launched a possession tracking tool that did the following:
  • Separated pace into its offensive and defensive components
  • Reported both pace and efficiency based on how a possession began. Nearly all possessions begin with either an opponent made shot, a defensive rebound, or a turnover.
For some time, this was the only place you could find pace split by offense and defense. The traditional box score definition of pace is unable to provide such a split. For a while, this split was available for the WNBA, but has since been removed. Today however, you can find offense/defense pace (along with pretty much any NBA stat you can think of) at pbpstats.com.

Pace can be a somewhat slippery concept. The traditional box score measure works well enough, but can also be affected by several factors that one would not normally consider part of "pace". For one, pace is largely a feature of team offense. But the box score definition measures the average of offensive and defensive pace. The 2014-2018 Warriors never got enough credit for how fast they played on offense because their strong defense led to consistently longer possessions from their opponent.

As mentioned above, possession length also depends on how a possession begins, with possessions following made shots being the slowest, those following live ball turnovers the fastest, and possessions following defensive rebounds somewhere in between. For a deeper dive into this concept, refer to Seth Partnow's recent article at The Athletic (with an assist from the lyrical majesty of Semisonic).

And possession length can also be affected by how often a team extends their possession via offensive rebounds. Possessions with an offensive rebound tend to last 6-7 seconds longer, but is not a quality we normally associate with pace.

Well actually, it's 11.5 seconds or less

As a complement to the efficiency and pace tool, I have created a tool to provide an alternate, and perhaps more direct, view of pace: How quickly does a team get a shot off after their possession begins? 


The Mike D'Antoni-era Phoenix Suns were known as the "Seven Seconds or Less" Suns, as their goal was to get a good shot within 7 seconds of starting their possession. According to this new tool, that looks like it was a stretch goal. In their heyday from 2004-05 to 2006-07, the Suns took an average of 11.5 seconds to get a shot off.

Like the possession tool, we can also split this stat by how a possession begins. A defining feature of D'Antoni coached teams is how fast they play after an opponent made shot. These tend to be the slowest possessions overall, but this is where the Suns played their fastest (relatively speaking). Despite their "fast break" reputation, the Suns were never a particularly fast team off of live ball turnovers.

How well does this measure of pace correlate with the possession-based view? The chart below compares average time to shot to seconds per possession (on offense).


As you can see, the correlation is fairly tight, with some deviations. The Bucks lead the league in offensive pace, but it's Alvin Gentry's Pelicans (Gentry is a member of the Mike D'Antoni coaching tree) that lead the league in how quickly they get a shot off.

Because we're not dealing with possessions as our fundamental unit, the tool also tracks how quickly teams get shots off after an offensive rebound. So far this season, teams take an average of 3.4 seconds to get a new shot off after rebounding one of their misses. The Pistons currently lead the league in this (rather volatile) stat, at 2.9 seconds.

With offensive rebounds added, the table can get pretty crowded. By default, Effective Field Goal Percentage (eFG%) on shots is hidden, but can be displayed by clicking the "Show eFG%" checkbox. Note that this table only tracks instances when a team gets a shot off (turnovers, fouls, jump balls, etc. not included).

At the bottom of the table, correlation between time to shot and effective field goal percentage is shown for each category. Similar to possession-based pace, a quicker time to shot generally correlates with better offense. Which raises an interesting question: do faster offenses create better efficiency? Or do efficient offenses lead to quicker shots? This may sound like a cop out, but I suspect it's both.

Just one final note regarding data quality: Starting with the 2017-18 season, I noticed that the average time it takes for a shot to be rebounded (according to the play by play data), increased by about 1.5 seconds. This increase appears to be a quirk of the NBA's new timekeeping systems, and does not reflect play on the court. And based on my spot checks, this appears to be a change for the worse. Rebounds are frequently recorded 1-3 seconds after the player clearly established possession. So be cautious when using this tool to compare pace across seasons. The average time to shot following defensive rebounds dropped from 11.7 seconds in 2016-17 to 9.9 seconds in 2017-18.

Powered by Blogger.